Odonatologica 9 (I): 75-99 March I, 1980

Territorial behaviour of the African libellulid

Orthetrumjulia Kirby (Anisoptera)

M.J. Parr

Department of Biology, University of Salford, Salford M5 4WT,

England

Received August 6, 1979

The the observations were carried outona colony of pale montane form of O.j.

falsum Longf. (8 April - 4 June, 1977) at Naisi, Zomba Mountain, southern

Malawi. 277 dd and 13 99 were given unique numbers on the wings which enabled

recognition without subsequentrecapture, and the positions of territorial dd were

recorded each the The accurately time colony was visited. most frequently seen

individuals showed the greatest consistency oflocalization. This indicates that the

most frequently seen individuals were superior in territorial defence and hence

tended to return most often toa chosen area. Whereas a territory might measure up

to 10 m in diameter, the central zone, where the spent 75% of its territorial

2 time, would be only about 2.5 m or less. Many territories were much smaller than ,

5 m in diameter. The outer zone territories of neighbouring often

overlapped. In the late afternoon, the highly structured territorial system broke

down and many widely ranging dd were observed to invade the colony area. In the

mornings the territories tended to be occupied assoonastheywere in partialor full

sunlight. The territorial dd spent 5-13% of their time in flight, depending on

weather About all conditions. 60% of territorial flights were aggressive and 20%

were feeding flights. 99 occurred at the breeding sites at any time during the

territorial period, but observed matings took place 10.13-14.45 hrs. The mean

copula duration was 29.5 sec. Usually the 9 oviposited in the d’s territory

there and her. immediately after copulation was attended by him, hovering over

for observed The mean survival mature dd was 9.0 days and maximum survival

was 51 days.

INTRODUCTION

It has long been recognized that behavioural studies of are

fruitful in providing valuable information about this taxon, in addition to

of of fundamental of contributing to our understanding some the concepts 76 M.J. Parr

behaviour. In particular, studies of territorial behaviour in the

Odonata, the and especially Calopterygidae, are increasingly

important in elucidating the complex relationship between localization

(residentiality), aggression and reproductive behaviour. Such studies of

specific taxa may prove to be of value in understanding analogous behavioural in animal patterns other groups exhibiting territoriality. The

work of SAKAGAMI et al. (1974) is an important attempt to begin the

analysis of the inter-related and variable nature of territorial and

behaviour in the Libellulidae reproductive using an evolutionary approach. The author considers present that behavioural studies of congeneric and other related species may be especially valuable in contributing to the

understanding of detailed taxonomic relationships (HEYMER, 1969;

PARR, 1973; PARR & PARR, 1974). For example, the eco-behavioural work the African on libellulid, Nesciothemis nigeriensis Gambles (PARR & demonstrated PARR, 1974) that this species, while conforming in most

libellulid respects to general behaviour, was apparently unusual in

maintaining mini-territories at the roosting sites, where reproductive behaviour was observed. It is only by a series of detailed studies of the

behaviour in natural surroundings that the full of libellulid range response will patterns be recognised and interpreted. However, the field worker

generally has to seize any available opportunity to investigate the detailed

behaviour of a given species as and when they present themselves.

Circumstances very rarely allow exactly parallel studies of related species to

be made in a planned sequence.

At the present time, about 25 libellulid species (representing 16 genera)

have been studied in detail with some particular reference to territoriality.

The best known in this include species respect Plathemis lydia (Drury)

(JACOBS, 1955; CAMPANELLA & WOLF, 1974), albistylum

speciosum Uhler (1TO, I960; ARAL 1972), O. coerulescens ( Fabr.) and O.

brunneum (Fonsc.) (HEYMER, 1969), O. cancellatum (L.) (KRONER,

1977), Crocothemis servilia Drury (HIGASHI, 1969), Pachydiplax

longipennis Burnt., (JOHNSON, 1962), Nesciothemis nigeriensis Gambles

(PARR & PARR, 1974), Leucorrhinia dubia (Vander Lind.) (PAJUNEN,

1962a, 1962b, 1963), L. rubicunda(L.)(PAJUNEN, 1963, 1966), L. caudalis

Charp. (PAJUNEN, 1964), L. pectoralis Charp. (KIAUTA, 1964), Perithemis tenera (Say) (JACOBS, 1955), Acisoma panorpoides inflatum

Selys (HASSAN, 1978) and Sympetrum parvulum (UfiDA, 1979).

The of Orthetrum population julia on Zomba Mountain, , was selected for the observations because the present colony was clearly limited in distribution which to a tree nursery was surrounded either by Pinus

or seminatural forest. plantations Furthermore, the were numerous and the site was easily accessible by landroverfrom Zomba town. Territorial behaviour of Onheirum Julia 77

As far as is known, the behaviour of O. Julia has not been investigated

main aims of this information previously. The study were to gather on localization of individual insects and the range ofactivities of males while in territorialsites. In addition some information was obtained on longevity and survival rates of males.

METHODS

The of direct Between 8 population of Orthetrum julia was studied by means observations.

April and 30 May 1977, totals of 277 males and 13 females (all post-teneral) were captured and

Observations marked with a unique number using orange or yellow enamel paint on the wings.

consecutive for the 8-12 18-24 were made on days periods April, April and 30 April-2 May, and continued intermittently until 4 June when the population had been observed for a total of 18

the days. As far as possible, observations commenced in the morning, as early as weather conditions would allow and continued for periods ranging from about I '/4 hours to nine hours, with of 3 hours 12 From the the work 8 and a mean minutes. commencement of on April, up to including 22 April, as many unmarked O. Julia as possible were captured and given a unique number, and the position of each marked male was recorded accurately on a plan of the colony area. Small numbers were also marked on 24 and 30 April, primarily in areas usually occupied by

marked On 24 and I males observed the already insects. April May individual were throughout

day. in order to record detailed behaviour within their defended territories. Times of arrival at territories were noted on three occasions and reproductive behaviour was recorded whenever seen. An estimate of male longevity was obtained from the resighting data of individuals marked

on the first eight sampling occasions.

THE HABITAT AND ODONATA

known Naisi Mountain The habitat was a tree nursery in the area as on the slopes ofZomba in

southern Malawi. Zomba Mountain (2087 metres), more accurately described as a plateau, has its western face clearlydefined by a great scarp along the fault line of the southern section ofthe

East African rift valley system. On the other side of the mountain, including the south-eastern area of Naisi, the scarp is less impressive but is also well defined with a vertical difference in

between the and the 1000 altitude surrounding plain plateau rim of approximately m. Naisi

nursery (Figs. 1-2) was established at an elevation of 1220m on one of the numerous flatter

from the mountainside. The shelves between sloping spurs projecting nursery occupied an area

measuring about 78 m by 72 m and was surrounded on all sides by forest. To the south-east the

site into drops away a steep-sided-ravine supporting semi-natural forest including Brachystegia

all spp. and Oxytenanthera abyssinica ( bamboo) adapted to constantly moist conditions. On

sides, site surrounded the of exotic other the nursery was on steep slopes by plantations pines,

A dominated by mature Pinus patula and P. kesiya. permanent, but very small stream flows

the the several through centre of nursery; this stream receives water from seepages originatingin

the itself. In the bed had been small nursery area seven places stream dug out to create pools to

facilitate collection of water for irrigation ofthe quarter ofa million or so tree seedlings. Much of

the surface ofthe nursery was terraced flat soil, but rank marsh vegetationwas present alongthe

stream and adjacent to the forestry road.

It was noticed that weather, in conjunction with the close proximity of a

high over-shadowing mountain had a distinctly limiting effect on the

activities of O. Julia at the Naisi colony. Cool, cloudy, misty and wet 78 M.J. Parr

conditions insects were common, so that on many days the would spend long

in the Weather in Malawi periods inactively, especially morning. patterns are complex because of the influence of mountains, the very large Lake Malawi and the Indian Ocean. However, although the main rainfall is limited to the

months of November - April, persistent low cloud, drizzle and low

of weather and windward temperatures chiperoni over rising ground facing escarpments may be experienced at any time of the year. 1 n the present study.

chiperoni weather predomin- ated during the period 9-11

April and rain or drizzle ended dragonfly activity on each of those days. The maxi-

mum temperature recorded at

the colony site ranged from

18.3°C on 11 April in chiper- oni conditions, to 26.1°C in

clear weather on 22 April. The

mean maximum for the study

period was 22.6°C. O. julia

flies the in throughout year

the Zomba area, although it is

less common during the cool,

dry period June - August.

Orthetrum julia (Fig. 3),

whicn was perhaps trie most

frequently seen anisopteran at

the Naisi nursery, has been

the subject of a certain

amount of taxonomic confu-

sion. It now seems certain

(LONGF1ELD, 1955; PIN-

HEY, 1970) that Othetrum julia Kirby, 1900, and O. fat-

sum Longfield, 1955, are con- I. Zomba Mountain Fig. Map of Naisi nursery area, with specific, and together Malawi, April-June. 1977.

their main variants and O. fal-

to sum capicola Kimmins, 1957, form a dineof races from West .

The very dark Orthetrumj. julia Kirby wouldappear to be the dominantform

in the dense forests of West Africa and the Zaire basin, but eastwards and

southwards much paler O. juliafalsum Longfield are found and O. julia capi-

cola Kimmins occurs only in south-west Cape Province of the Republic of

at Naisi to with South Africa. The specimens of O. julia seem agree best the Territorial behaviour of Orthetrum Julia 79

2. Naisi Fig, nursery, Zomba Mountain, looking south-east

pale montane form of O.

julia falsum described by

PINHEY (1970).

In addition to O. julia,

nine other Odonataspecies

were seen at Naisi nursery

the observation during pe-

riod. Of these, Crocothe-

mis sanguinolema (Bur-

meister) was most com-

mon and it established ter-

ritories in the same areas as

O. Julia. Small popula-

tions of Orthetrum abbotti

Calvert and Crocothemis

of male Fig. 3. Photograph a Orthetrum julia, at Naisi. divisa Baumann existed

for the duration of the

and numbersof kersteni study period, large Pseudagrion (Gerstaecker) were associated with the stream margins. A few individuals of Orthetrum caffrum

(Burmeister), Palpopleura lucia portia (Drury), P. jucunda (Rambur),

Diplacodes lefebvrei (Rambur) and Pantalaflavescens (Fabricius) were seen the at nursery but without any evidence of breeding. 80 M.J. Parr

LOCALIZATION

RANGE OF OBSERVED LOCALIZATION

In the territories colony studied, most were bare or relatively bare areas of

but not soil, often, invariably, immediately adjacent to a stream or seepage.

The in territory was normally only occupied consistently sunny weather and favoured most perches were small sticks projecting from the surface, stones or rocks. Some individualsestablished territories amongst waterside vegetation and then rested on tall stems, or grasses overhanging the water.

In common with many other dragonflies exhibiting territoriality, it was noticed that O. julia appeared to show a tendency to return repeated-

ly to the same small area. The detailed analysis of the posi- tions of all marked males in the colony has allowed some conclusions to be reached regarding the predictability and constancy oflocalization.

The information plotted on the site each time map the colony was visited (e.g. Fig. 4) allowed the position(s) of each male O. Julia to be marked on individual site maps (selected individuals’ positions are shown in Fig.

5a-1).

Table 1 the numbers gives Fig. 4. Positions of territorial male Orthetrum juliaat of males seen two (62 insects) Naisi on 23 April, 1977. (cf. Figs. 1-2). or more times, up to a maxi- mum of 14 times (one insect). Column 3 of Table I gives the mean maximum

of times number individuals were localised within a circle of 10 m diameter, and col. 4, obtained from the ratio col. 3/col. 1, gives the 'localization index’

(I). The parameter I is a measure of individuals’ mobility, within the context of territoriality, related to the number of times seen. The circle of 10 m diameter was chosen to measure localization because most individuals showing a tendency to return to a given spot were apparently responding to

landmarks within of recognized an area approximately that size or less. Territorial behaviour of Onhetrum Julia 81

5. Outline of Naisi of males of Fig. maps nursery showing positions selected Orthetrum juliaon

each occasion they were seen (cf. Figs. 1-2). The ringed numbers indicate the individual was seen that number of limes within the 10 diameter circle. m Single dots indicate the individual was seen oncein that No. 6. 14 times. Localization index — 11 position: (a) seen (I)=0,93; (b) No. 131, seen — = = times. 1=0.91; — No. 34. seen 9 times, 8 1 (c) I 0.78; (d) No. 144. seen times, 0.88; — (e)

8 — No. 7 No. 100. times. I =0.50; 20, times, I — seen (f) seen =0.57; (g) No. 32, seen 6 times, — — I = 0,33; No. 52. seen5 times, 1 =0,40; No. 103, times. — (h) (i) seen 4 I =0.00; No. 221, seen

4 times, I = = — 1.00; (k) No. 260. seen 3 times, I 1.00; (1) No. 258, seen 3 times, I =0.00. 82 MJ. Parr

Table I

Calculation of localization index

y 2 3 4 5

times times Localization No. No. insects Mean max. no. Standard

seen in sample localized within index (I) deviation

10 m circle (Col, 3/Col. 1) ofl

14 1 13.00 0.93

II 2 9.50 0.86 0.045

9 4 7.00 0.78 0.078

8 7 5.57 0.70 0.188

7 7 3.71 0.53 0.144

6 9 3.00 0.50 0.158

5 22 2.09 0.42 0.255

4 32 1.75 0.44 0.354

3 51 1.18 0.39 0.384

— 2 62 0.23 0.12

It was evident from analysis ofall the datathat few individuals were wholly

in territorial circle. it did predictable always returning to a 10 m However,

seem as if the most frequently encountered individuals were generally more predictable in where they might be found. In order to remove possible bias by recording static individuals relatively more frequently than mobile ones, the colony area was searched systematic- ally each time it was visited, so that this probable source of error was minimised. If there were no relation- ship between the number oftimes an individual was seen and the probabil- ity of its repeated appearance in the

same territory, the localization in-

dices would not be expected to rise steadily as the number of sightings increased. However, inspection of

Figure 6 and Table 1 shows that, in

fact, the parameter 1 is very signific- Fig; 6. Orthetrum julia: Correlation between antly higher in frequently seen insects Localization lndex(l)andnumber oftimes seen than in those seen few times. The cor- 0.0637x + (r= 0.955, p< 0.001; y= 0.1274). relation between the localization

is with = for index and number of times seen (Fig. 6) strongly positive, r 0.955

d.f., line of on x in 6 is based on 9 p < 0.001. Although the regression y Figure a linearrelationship between 1 and the numberof times individualswere seen, Territorial behaviour of Onhelrum Julia 83

an exponential model is probably best used to describe the relationship

is between these two variables. The value of 1 for individuals seen only once

maximum value is set at zero and this parameter’s possible 1.0, when the

is is statistical number of times seen infinity. It hoped to investigate the

of localization index later date. aspects the at a

MICRO-LOCALIZATION AT THE TERRITORIAL SITE

The most frequently seen individuals have been shown to have very

definite preferences for restricted areas within the colony boundary.

However, inspection of Fig-

for ure 5b-d, example, shows

this is that usually so, but the

small scale of the maps ob-

scures the fact that the territo-

defended ry on any particular

day is apparently slightly dif-

ferent from other days. In

order to investigate the ter-

ritorial localization of a par-

ticular individual on a given

day, male no. 144 was kept

under constant observation

This individual on 1 May.

(Fig. 5d), was seen seven times

in out of eight the same small Fig. 7. Diagram of territorial area showing micro- this oval area. On 1 May male localization of male Orthetrum julia, no. 144 on I May, first at09.15 and hrs individual of its time appeared 1977 (cf. Fig. 8). The spent 75% 15.45. this departed at During within the inner zone (heavy broken line) and 25%

within the outer zone (light broken line). The dots and period of exactly six hours a numbers indicate the of time its percentage spent at a record was kept of exact particular perch within the territory. time position and the spent on

each perch. The individual

perches selected by this maleare marked on Figure 7 with dots and the figures

denote the percentage time (out of six hours) spent at each perching site.

Except in the course of some of the longer territorial defence flights, this

individual spent the whole ofthe six hours within the territorial boundary of

about 7 mx 5 m shown on Figure?. However, 75% ofits timewas spent within

the oval central zone of approximately 2.5 m x 1.0 m. The territory was

divided it noticeable by a small seepage stream and was that most of the

insect’s time the bank. It be from 8 that was spent on right may seen Figure

much of the territory of this insect was nearly bare soil with a very thin. 84 M.J. Parr

Within this the favoured discontinuous, grass cover. area most perches were

small sticks lying on or projecting from'the soil surface and the white rock

visible in Figure 8.

Not only did the length of

time each spent at position

vary greatly (Fig. 7), but also

the insect spent a number of

separate periods of varying

lengths of time in each perch-

ing position (Tab. II). For

example, during the day it

defended its territory for a

total of40 min. from position

2 (11.1% of the 6 territorial

hours), comprising 8 separate

periods averaging five min-

utes each. Within each period

at a particular perch the insect

made a varying number of

flights (mainly feeding and

aggressive): if it settled on a

different perch at the comple-

tion of a flight this was re-

corded. It may be seen that

the dragonfly tended to return

most frequently to those

perching positions (2, 5,6 and

which 8) at most time was

spent (63.6%), and that it ex- Fig. 8. Photograph of territory of male Orthetrum hibited strong preferen- The very julia, no. 144 on I May, 1977 (cf. Fig. 7). territorial for certain and the ces perching sites boundary is marked with a broken line territor-

with line. ial inner zone is marked a continuous over other possible ones.

These observations on I

May suggest that many male individuals are, in fact, more strictly localized

when the same area is defended on successive occasions, than at first appears

to be thecase. For example, no. 131 (Fig. 5b) was defending the same territory

on 10 out of the 11 times it was seen, although its precise perching site within

that area varied. No. 6 (Fig. 5a) was recorded at almost exactly the same

Colocasia position (a small group of coco-yams, esculenta) on 12 of the 14

occasions it was seen. This male was unusualin that it habitually rested on the

leaves of these plants, and not on other vegetation or the ground.

In most cases the outer zone territories of neigbouring male O. julia Territorial behaviour of Onhetrum Julia 85

Table II

Micro-localization at territory site: No. 144, 1 May, 1977

I 2 3 4 5

No. Mean & Perching Total time (min) % of time (6 hrs) of indiv. range

position spent at each spent at each periods spent of length of

(Fig. 7) position position at each position period (min)

(Col. 2/Col. 4)

1 16 4.4 1 16.0 (16)

2 40 III 8 5.0 (1-16)

3 9 2.5 2 4.5 (1-8)

4 6 1.7 1 6.0 (6)

5 103 28.6 8 12.9 (4-21)

6 38 10.6 3 12.7 (1-24)

7 3 0.8 2 1.5 (1-2)

8 48 13.3 5 9.6 (1-17)

9 20 5.6 5 4.0 (1-10)

10 1 0.3 1 10 (1)

II 4 l.l 2 2.0 (1-3)

12 17 4.7 1 17.0 (17)

13 10 2,8 2 5.0 (2-8)

14 25 6.9 2 12.5 (10-15)

15 8 2,2 2 4.0 (4)

16 4 l.l 1 4.0 (4)

17 3 0.8 2 1.5 (1-2)

18 5 1.4 1 5.0 (5)

total 360 99.9 49

to considerable but it to overlapped a extent, was rare see any other male

being allowed to remain within the inner zone, although this did sometimes

size territories happen. The of varied in different parts of the nursery and in

several areas the territories were often much smaller than five metres in

diameter (Fig. 4).

life A table derived from the marking and resighting data gives a mean observed survival of 9.0 days for mature male O. julia. The maximum observed survival times were 51,49 and 40 days by males no. 34 (Fig. 5c), no.

103 and 166 These survival times (Fig. 5i) no. respectively. can probably be increased minimum by a of 10 days to account for the maturation period of

the young dragonfly.

ACTIVITY PATTERNS AND TERRITORIALITY

POST-TERRITORIAL, ROOSTING AND EARLY MORNING BEHAVIOUR

Male O. their territories Julia abandoned on most days between 15.00 and 86 M J. Parr

their in 16.00 hrs, depending on position the nursery (some areas stayed sunny longer than others) and weather conditions. In very bad conditions such as chiperoni weather, the population would not appear at the territorial sites at all. Sometimes individual would shift its another of an position to part the nursery when its territory became shaded in mid or late afternoon. This resulted in some well-established territory holders adopting short-term new territories in sunny sites at the end of the afternoon, prior to leaving the area for time occasional roosting. At any during the day one could see male individuals in nomadic fashion and associated behaving a very clearly not with definite wanderers, These individuals any territory(= HIGASH1, 1969). were usually unpruinosed on the thorax and spent much oftheir timein active flight. It was presumed that they were young insects. It was noticeable that the numbers of such peripatetic (usually unmarked) individuals increased towards the end of the afternoon when some territory holders were changing their positions. It is assumed that many of these very mobile males were unsuccessful in establishing normal territories. This was either because they

older unable with were young, or were and to compete successfully established males because of other, not necessarily age related, factors. In the late afternoon, the closely structured system of territories broke down and this presumably allowed other, less territorially successful males to remain if they flew in from surrounding areas. During this period of the day. before

of roosting commenced, the numbers of males was high, a great deal flying

aerial of activity took place and complex combats involving groups males (10 on one occasion) were frequent. For example, on 1 May, when male no. 144 was kept under close observation until it left its territory at 15.15 hrs, numerous flying and hovering unmarked males were present between this time and 15.45. These males were very aggressive and were patrolling areas where a few minutes previously well established territory holders such as nos.

30, 131 and 144 had been in residence. All of this late afternoon activity at

Naisi nursery took place in shadeand in a temperature of about20.5°C. with

of afternoons, 16.00 hrs few a relative humidity 81 percent. On most by very

O. julia could be seen flying. Roosting appeared to occur in the Pinus plantation canopy, since males were seen to fly into the tree tops on leaving the open nursery area. However, because of their inaccessibility, it was impossible to confirm this observation.

of in determined The appearance male O. Julia the mornings was mainly by the presence ofsunshine. On sunny mornings males could be seen flying out of the plantations on the west side of the nursery from about 07.15 hrs. An

it in full individual’s territory tended to be occupied as soon as was partial or

in of Orthetrum coerulescens and O. brunneum sunlight, as the cases

(HEYMER, 1969). For example, on 2 May, no. 122 was in its territory at

3 267 08.05, 07.30, no. 34 at 07.39, no. 131 at 07.43, no. 67 at 07.56, nos. and at Territorial behaviour of Onhetrum julia 87

271 30 no. at 08.11. no. at 08.16 and no. 6 at 08.17 hrs. These times approximate closely with the times the territories first received sunshine that morning. When no. 131 was first noticed at 07.43 it was attending an ovipositing female when the site was just in sunlight. When individuals first arrived their exhibited at territory they only weak attachment to the area and were readily disturbed. On 2 May, nos. 131,67 and 6 flew back to the edge of the nursery when they were disturbed soon after their arrival and perched there in sunshine for some minutes before returning to their territories. Later in established difficult the day, well males were to displace by moderate amounts of disturbance by human activities such as movements of forestry workers or myself. On cloudy mornings or in areas receiving sun relatively

late in the day, males arrived very much later than the foregoing times.

time Females would appearand oviposit at any during the period of normal activity of males, but the recorded matings were all between 11.40 and 14.45

hrs.

TERRITORIAL FLIGHT ACTIVITIES

On two days detailed observations were made on individualsfor the whole

of their territorial periods. Males no. 20 and 131 were observed from 10.51 -

11 .00 and 11.27 - 15.16 hrs respectively (Tab. Ill) on 24 April; and no. 144

from09.15 - 15.45 hrs on I May (Tab. IV). A preliminary classificationofthe

types of flight and their characteristics in territorial male O. julia is given in

fable V.

On 24 April, no. 20 was abandoned at 11.00 hrs because its particular

territorial site made observation difficult, and no. 131 was observed for the remainder of the The reached 07.20 day. nursery was at hrs when there was

total overcast and mist was drifting across the colony area; the dry bulb/wet bulb temperatures were 18.3°C/ 17.7°C. The first male O. julia to appear flew

down from forest 07.25 the at hrs, crossed the nursery and disappeared. In the

07.25-07.50 several male Crocothemis period sanguinolenta were active by

the stream and small ponds, but no more O. juliaappeared until07.51 when

32 tall the ofthe in no. was seen resting on grasses at extreme edge colony area

very misty conditions. Another marked male appeared at 07.55 but flew off before it could be identified. At 08.30 (I7.7°C/ 17.5°C) it was raining lightly and rain continuedintermittently until 10.06 hrs. From this timeonwards the

weather conditions gradually improved and the sun appeared for six minutes

at 10.41 hrs, which resulted in more male O. julia flying over the nursery. No.

20 arrived its 10.47 at territory at and immediately the sun was obscured by

cloud. However, when the observations were transferred to no. 131 the

weather became and considerably brighter dragonfly activity stayed at a high

level until observations on this individual ended at 15.16 hrs. 88 M.J. Parr

5.69 0.81 (?) 15.42 4 5 30/1 13 (30.8) 4 15.01-15.16 (38.5) 18.9/18,3 100 no.20,

male 6.69 1.20 (?) 38/1 13.39 5 14.31-15.00 36 (13.9) 27 (75.0) 4 20.5/18.9 83 to refers 3.87 1.27 8.17 (?) 12/1 6 38 (15,8) 25 (65.8) 7 83 column 14.01-14.30

first (sx) 7.41 1.23 1 15.22 9 86/1 37 (24.3) 22 (59.5) The 13,31-14.00 (?), 23.3/20.0 100

5 1977. (?)

1 1.60 1 5.17 13.78 5 47/1 48 (10.4) 31 (64.6) April, 13.01-13.30 (a), 100

131. 1 24

on no. (?) 19.4 4 111 5.36 0.93 8.33 19/1 5 18 julia male 12,31-13.00 28 (17.9) (64.3) (a), 23.3/ 100

1 Table to

(?)

others 8 Orthetrum 1.43 9.3) (c) 8.16 19.50 4 45/1 43 27 (62.8) (st), I 100 the 12.01-12.30 ( male 3

(?) two 131 1.05 3 (sx) 20.75 5 of 11.86 51/1 21 (23.8) II (52.4) No, 11.27-12.00 (st), 1 22.2/19.4 93

1 activities 20 18.9 0.62 3.59 (st) 3.50 8/1 1 8 5 0 (62.5) (12.5) 2 flight No. 10.47-11.00 20.0/ 6.5±9.2) territorial % % sexual, period mean and and copulatory; of = undeter- of each fits each sx of flight = end duration 1.2) 13.3) flights = c (Overall in in 17.6) 61.4) flights ? sec flights/min at Summary period = = sunny sec Min. in mean mean spent flights feeding % aggressive % other copn.; duration no. stabilizing; temp, in & avoiding; time flights of of of of time = of Mean flight Max. Mean (Overall (Overall specified (Overall (Overall without = mined) Shade of % No. No. No. No. (st a % Territorial behaviour of Orthetrum Julia 89

It be from Table III that the oftime may seen proportion spent inflight by

nos. 131 and 20 was relatively low, ranging from about four to 21% in

different periods. The mean duration of each flight was 6.5 ± 9.2 sec. Apart

from sexual all one flight, flights lasting 10 seconds or longer were aggressive

in the 86 is noticeable nature, longest being sec. It particularly thatapart from

the first period in the day, which was sunless throughout, aggressive flights

were much more numerous than flights of any other type. The overall

of with percentage aggressive flights was 61.4%, compared 17.6% for feeding

the flights, next most frequent. Aggressive flights were mostly directed at other male O. julia and could be easily recognized, being essentially similarto

those described for the West African libellulid, Nesciothemis nigeriensis

(PARR & PARR, 1974). Three ofthe aggressive flights made by no. 131 on 24

directed April were at male Crocothemis sanguinolenta, which were probably

O. as numerous as Julia. Normally, C. sanguinolenta were ignored by male O.

Julia even when they were perched or flew within a few centimetres of the

territory defender. Feeding flights were invariably quite short (nearly always

less than eight seconds duration) and were spaced relatively regularly through

the of undetermined day. Flights nature were recorded throughout much of

the day. In most cases these resembled feeding flights but the dragonfly did

not catch any prey and no prey target could be seen by the observer. It seems

likely that these were unsuccessful feeding flights and that the prey were very

small insects. In verified either feeding flights, the prey was seen by the observer before being caught by the dragonfly, or masticatory movements of

insect’s the mouthparts were visible. Stabilizing flights occurred, forexample, if insect the was momentarily unbalanced by a gust ofwind. On one occasion,

sudden of disturbed a movement the observer the insect, which made a one-

second before Three sexual flight settling. flights were made by no. 131 on 24

Two chases of females without April. were contact being made; a third was a

four-second unsuccessful copulation attempt.

During the second all-day observation of a territorial male O. Julia (no.

144), the weather variable with was sunny and dull periods alternating (Tab.

In this individual IV). general, was less active than was no. 131 on 24 April.

The overall of time in 5.2 percentage spent flight was and the mean duration

of each flight and mean number of flights per minute were much lower than

for 20 131 combined. nos. and There is a significant difference in the mean

duration of each between the — flight two dates: nos. 20 and 131 combined,

24 6.5 ±9.2 I 4.9±4.1 for April; sec; no. 144, May; sec, which d =2.5213,and

0.02 > p > 0.01. During the several overcast periods flight activity of O. Julia

males was suppressed and this factor alone is sufficient to explain the

difference in overall levels activity between 24 April and 1 May. However, a

similar of very proportion feeding to aggressive flights occurred on the two

occasions: 20 and nos. 131, 17.6% feeding, 61.4% aggressive; no. 144,24.4% 90 M.J. Parr

•• 6.46 24/1 0.53 5.74 (?) 3 14.31-15.15 24 (12.5) 21 (87.5) 1 100

(a) 4.09 1 0.70 4.78 1 7 14.01-14.30 15/ 21 (33.3) II (52.4) (?). 33 2

(st) 3.00 0.27 1.33 1 7/1 8 4 (50.0) 1 (12.5) 0 13.31-14.00 (?), 21.1/18.3 2

0.47 2.83 (?) 3.64 8/1 6 6 13.01-13.30 14 (42.9) (42.9) 2 13 julia

(P) 6.35 1.13 1 12.00 8 28/1 34* 24 90 12.31-13.00 (23.5) (70.6) (?). 22.8/17.8 Orthetrum 2

(a) male 4.15 0.43 3.00 1 12/1 4 5 17 12.01-12.30 13 (30.8) (38.5) (?).

3 twice. 00 (st)

territorial 1 1 IV 5.17 0.77 6.61 (P) twice. 3 14/ 23 (13.0) 16 (69.6) 1 87 scored 11.31-12. (?), 22.8/18.3 in Table 2 scored (sx) therefore

flight 1 7.89 5.68 0.83 is 17/1 25 4 (16.0) 19 (76.0) of 11.01-11.30 (inv), 100 therefore and

1 is types a, (a) and as

3 (P) of 3.76 0.83 5.22 a, 16/1 7 0 25 (28.0) 12 (48.0) 1 10.31-11.00 (?) 22.2/17.8 as ended 2 ended but (inv) but 0.67 5.06 1 4.55 5 feeding 14/1 20 9 67 Classification 10.01-10.30 (25.0) (45.0) p (?). as as 5

(a)

4.61 0.40 3.07 1 20/2 4 9 commenced commenced 09.15-10.00 18 (22.2) (50.0) (?). 22.8/18.3 22 4 4.9±4.1) which which % % sexual, period sec sec 28 13 mean and and = of cacn = a undetermined) of of each fits investigatory; sx end of duration flights/min 0.6) 5.2) flights 24.4) 59.1) flight flight in flights = = (Overall flight at

sec = = ? sunny 11 of in period copulation; inv sec Min. mean mean duration in flights feeding % aggressive % other stabilizing; temp, no. in & time patrol; time of of of of Includes Includes flight of = of avoiding; = Mean flight Max. Mean (Overall (Overall No. specified No. (Overall No. (Overall No. without Shade of % (st p % •• •• Territorial behaviour of Onhetrum julia 91

feeding, 59.1% aggressive. Other less detailed observations underlined this fact that in the Naisi colony of O. Julia intraspecific aggression was the , dominant form of flight activity during the territorial periods. One of the aggressive flights on I May was directed at a male Crocothemis sanguinolenta. Other flights performed by no. 144 included 25 of

undetermined nature (probably mostly unsuccessful feeding attempts) and

which Six one sexual, was non-copulatory. avoiding flights were caused by

disturbance, such as the movement of a bee nearby large or wasp

(Hymenoptera), a lycaenid butterfly (Lepidoptera) and once, another male O. julia. The latter situation is rare, since male-maleinteraction in O. Julia is

virtually always aggressive. On two occasions, movements of grasshoppers

(Acrididae) seemed to initiate investigatory flights which were different from

These of short-medium anyother flight pattern. investigatory flights were

length (5 and 12 sec), more or less horizontal and relatively leisurely in pace.

On I May no. 144 made three patrolling flightswhich differed fromall others

in being apparently spontaneous and with no definite orientation. The first

two were five and six seconds long and the insect flew in a roughly circular

horizontal course from its perch and then returned. On the third occasion a

similar flight was transformed into an aggressive one when a neighbouring

This male flew up and intercepted the patrolling specimen. resulted in a fast and extended chasing flight of 28 sec in all, which was the longest flight recorded for no. 144 that day. The function of such patrolling flights is

uncertain, establish but they may serve to the current distribution of

males. neighbouring conspecific However, they do not seem to have occurred

in the Naisi and recorded all 24 very commonly colony were not at on April.

REPRODUCTIVE BEHAVIOUR

seemed without female Mating always to occur any prior courtship, the

chased and seized frequently being immediately she appeared over the

territorial areas of the nursery. However, females were also seen, on occasions, ovipositing unattended, and apparently in view of territorial

males. females Although could be seen throughout the day flying over the

breeding sites, all matings took place between 10.13 and 14.45 hrs, but the

restrictiveness of these sightings is probably misleading. For example, on 2

May at 07.43 hrs, male no. 131 was seen hovering above an ovipositing

female, whose oviposition site was just in sunshine at that time. Since

attending flights always appear to be associated with the very recent mating of

the female concerned, this evidence for may be taken as an early morning

copulation. The female is seized in flight and commonly, the pair fly in

tandem for a few seconds before assuming the wheel position. When

copulation has commenced, the pair may continue flying for some seconds 92 M.J. Parr

distribu- of function territory or seen current (only mating on balance dd? of unsuccessful flights neighbouring 9 recent 9) defence regain be of Comments position Protective feeding Establishes conspccific Guarding In — To — — May tion after same

flying area or nearby spontaneous; after insect of <5 9 of julia behaviour 9 by balance perch territory prey of of Initiated Conspccific on Disturbance Appearance Movement Apparently copulation within Aerial Loss grip conspccific insect intrinsic pattern? Release Orthetrum ? male approach approach speed below direct 9 darting darting more medium cruising direct or direct and over territorial flight (or V and and of & side & above in Table Type Rapid from Rapid Fluttering Rapid Rapid from Leisurely Rapid Circular complex) horizontal Hovering flight

of so often being sec. types nearly oviposit- intruder prey, object 16+ of 9 upwards 9 or from territory = in Long Orientation Towards Towards vertical Away disturbance Towards Towards investigated Usually Indefinite Towards Classification — ing sec; =6-15 Medium Duration Short-Long Short Short Short Short-Long Short-Medium Short Short-Long Medium-Long

sec;

1-5 = — flight incl. Short of Undetermined : Type Aggressive Feeding Stabilizing Avoiding Sexual, copulatory Investigatory Patrolling Attending Note Territorial behaviour of Orthetrum Julia 93

before settling or may settle immediately.

Completed copulations of O. julia were observed on six occasions. These

lasted 11,21,28,29,39and 49 sec: mean time29.5± 13.3 sec. Thus O. julia falls

within the upperrange of short-type copulations (CORBET, 1962, p. 178). In

of female all but one these cases, the began oviposition attended by the male

immediately the pair separated. The female in the copulation lasting 28 sec

rested for 6 sec before ovipositing attended by the male. In the same habitat

Orthetrum abbotti was observed to mate for 6 min 55 sec and 9 min 36 sec

while settled. In both of these instances the female did not oviposit on

separation, but flew away and out of sight, in one case after resting for some

minutes. The males of O. abbotti made no attempt to stay with their mates

after separation had occurred. A complete mating of Crocothemis

sanguinolenta on 24 April lasted 7 sec.

DISCUSSION

it form localization In recent years, has become clear that the or

residentiality may take in libellulid dragonflies varies considerably. In.part,

some of these variations seem to depend on whether males respond differently

to the two sexes or whether all approaches initially are sexual (JOHNSON,

1964). In mature male libellulids, localization within a small area of a few

square metres always appears to be linked with territoriality. However, in

male Nesciothemis nigeriensis which show strong localization during

maturation, territorial behaviour is not in evidence until sexual maturity is

attained (PARR & PARR, 1974). The variable degree to which localization

be said male MOORE may to apply to dragonflies was recognized by (1957)

who pointed out that localization of territory is not complete and that there is

a gradation between unlocalized fighting, and fighting and displaying in a

clearly defined area. FALCHETT1 & UTZER1 (1974) make the observation

in Crocothemis defence that erythraea, only flying periods are periods; resting

spells occur outside the territory, so that there is discontinuous defence of

territory. The study demonstrated that in Orthetrum Julia some present ,

individuals are strongly localized at the breeding area whereas othersare not.

PAJUNEN (1964), KRONER (1977) and HASSAN(1978) have commented

on the same fact in their respective studies of Leucorrhinia caudalis,

Orthetrum cancellatum and Acisoma panorpoides inflatum. This seems to

involve at least two major factors. Newly matured males of O. Julia may have

in difficulty establishing a definiteresidential area and will then function as

’wanderers’. opportunistic Secondly, more experienced males may or may not

establish localization other factors which long-term depending on may collectively be termed ’reproductive fitness’. Those insects with relatively low

fitness' ’reproductive may join the inexperienced ’wanderers’ and tend to 94 M.J. Parr

appear anywhere in the breeding area. The study of Crocothemis servilia by

HIGASHI introduced the i.e. in (1969) concept of’wanderers’, this context,

mature males without territories, and suggested an important relationship

between the total number of males present and the number of territory

holders at a pool. HIGASHI (1969) and HEYMER (1969) also collected

evidence of libellulids that the residentiality male increased when a territory is

established. This principle has been seen to apply to the Naisi population of

O. Julia and a strong correlation between the localization index and the

number of times seen (Fig. 6) was demonstrated.This phenomenon is clearly

linked with the variable localization noted by several workers. It appears that

the most frequently seen individuals of O. Julia were superior in their defence

of a territory and, therefore, tended to return most often to a particular

chosen area. Unsuccessful males, because they were unable to compete

effectively with the dominant individuals, tended to return to the general

addition colony area less frequently, in to failing to defend a particular

territory for many successive occasions. These relatively unsuccessful

considerable of their time 'wanderers’, therefore, spent a proportion away

from the breeding sites.

Leucorrhinia In rubicunda, PAJUNEN (1969) reported that some males

persistently occupied certain sites, moved less than others and tended to be strongly aggressive, which lead to a regular spacing of males. Some

individuals of Plarhemis lydia also showed almost total localization

fact in (JACOBS, 1955). The that dense populations there was a contraction

of size with territory compared sparse populations was emphasized by

KORMONDY (1961), HEYMER (1969) and KRONER (1977), and this does

be in other Odonata. There be direct seem to a general occurrence may a

relationship between size of the insect and size of territory defended, with the smallest species, such as Nannophya pygmaea (YAMAMOTO, 1968)

utilizing the smallest territories. The density of male O. Julia at the Naisi

nursery was high and most of the territories had boundaries that overlapped

to varying extents: for example, no. 144 on 1 May shared some of his outer

with zone territory numbers 8, 30, 131, 170 and 271. Although considerable

of territorial boundaries in overlapping occurred O. Julia, any particular

male’s dominance and aggression was clearly evident in the central zone he was defending.

in The territoriality demonstrated by O. Julia seems some ways to be

intermediate between that shown by O. coerulescens and O. brunneum

of surfaces (HEYMER, 1969). O. Julia perched on a variety and the territory

may be very open and flat soil or tall waterside vegetation such as grasses, whereas the othertwo species observed by Heymer were much more distinctly

polarised towards vegetation or open rock surfaces.

In considering the length of time a particular male dragonfly occupies a Territorial behaviour of Onhelrum julia 95

territorial of this is given area some cognizance must be taken how measured.

Some species may have individuals which repeatedly return to the same territorial area for several days but may not remain in the territory for the whole day, as in Leucorrhiniadubia'(PAJLINEN, 1962a). In L. dubia, several

utilize territorial different times in males may the same area, but at the day.

Male Plathemislydia were shown to be largely classifiableinto those arriving

at water in the morning (13%) and those arriving in the afternoon(60%) and

Perithemis tenera apparently behaved similarly (JACOBS, 1955).

JOHNSON (1962) recorded males of Pachydiplax longipennis as remaining

in their territory for periods of between 10 minutes and three hours. Males of

Nesciothemis nigeriensis remained in waterside territories for about five

hours each day (PARR & PARR, 1974) and Orthetrumjulia, in the present

study were territorial for up to eight hours per day (usually about six),

depending on the hours of sunshine. HEYMER (1969) established that male

O. coerulescens and O. brunneum also tend to remain in their territories

throughout the day.

The work of CAMPANELLA & WOLF (1974) has suggested that in

Plathemis lydia the system oflocalization and territoriality which has evolved

is different which described from any has previously been for the Odonata. In

the P. lydia habitat studied by Campanella & Wolf, males defended certain

traditionalareas for the primary purpose of mating. At any one timewithin a

particular territorial area two, three, or more males would interact

aggressively with each other to establish a dominance hierarchy and each

individual on the area appeared to recognize and maintain the territorial

boundaries. Thus, territories of P. lydia were not used exclusively by a

dominant individual. This involved the recognition of dominance and

submission by the insects using a particular area, so that the dominantmale

was most successful in mating attempts. Campanella & Wolf describe the

of territoriality and mating P. lydia as a temporal lek system where males

traditional site for return to a communal the primary purpose of mating.

Within the temporal lek system the most dominantindividuals command the

optimum timefor reproductive activities. The temporal lek system described

for P. lydia seems to be a complex evolutionary solution to the problem of

high density of male reproductive individuals competing for the limited

resources of space, time and females.

In contrast to the complexities of territoriality and localization in P. lydia,

the systems in Orthetrum albistylum speciosum (ITO, 1960; ARAI, 1972), O.

coerulescens and O. brunneum (HEYMER, 1969), O. triangulare melania

(ARAI, 1972), O. cancellatum (KRONER, 1977) and O. julia are relatively

simple in that mature males defend definite territories of their own. In O.

julia, each territory has a virtually inviolate central zone, although the outer

with of zone does overlap the outer zone other neighbouring territories. 96 M.J. Parr

Hence, in O. and in julia, presumably other Orthetrum species too, there is no

hierarchy of dominance, and no multiplicity of males occupying the same small Since females area. show no preference for visiting the breeding site at time of any particular the day, there is no optimal period for males to exhibit

in O. It is, therefore, that territoriality Julia. not surprising males, once established in a territory tend to remain there all day. The West African libellulid Nesciothemis nigeriensis conforms quite closely with the general of localization pattern and territoriality set by O. Julia. However, in N.

nigeriensis the situationat water is perhaps simpler than for O. Julia in that the

territories were and usually larger appreciable overlapping was certainly absent in the northern Nigerian colonies studied by PARR & PARR (1974). P. would lydia appear to be a ’percher’ that spends more of its time in flight

than O. or N. julia nigeriensis. Other libellulids such as Pantalaflavescens and

Tramea basilaris With are typical ’flyers’. such a wide range in flight activity

levels amongst the libellulids it is to be expected that localization and

territoriality patterns vary greatly.

Detailed of territorial analyses activity patterns have been made on

few Odonata. relatively In the present study of Orthetrum julia, and in the

work on Nesciothemis nigeriensis (PARR & PARR, 1974) individual males

under observation for were kept as long as they remained at theirterritories in obtain order to a clear of how the territorial is picture exactly period spent. From these studies two it is seen that territorialactivities are.much influenced

weather and that the by two species do behave similarly. It should, perhaps,

be pointed out that the Nesciothemis is assumed to be allied very closely Orthetrum to and that one might expect close similarities in behaviour.

However, the amount of time in the territory that is actually spent in flight differed in the two species (O. Julia about 13% in fine sunny weather, 5% in mixed weather with some sun and overcast periods; N. nigeriensis 8% in fine

in mixed These weather, 3% weather). seem to be genuine specific differences since active the more O. Julia lived under relatively cool and effectively temperate conditions in Malawi compared with the much hotterand sunnier climate experienced by N. nigeriensis in the northof . The proportion of feeding flights was quite different in the two species. In O. Julia about 20% of all within territorial flights a context were feeding flights and 60% were

directed males. In aggressive flights mainly at conspecific N. nigeriensis, over

60% were feeding and about 30% were aggressive flights. The difference in the rates of aggressive encounters is probably associated with the observation that in O. territories Julia neighbouring overlapped and hence the males were often closer together than in N. nigeriensis where territories were more individually defined. It is uncertain whether territory size and aggressiveness is reflection of in a simple population density O. Juliaand N. nigeriensis-. it is that possible each species has its own intrinsic upper and lower limits of Territorial behaviour of Orthelrum Julia 97

territory size. Quite often, the territories of N. nigeriensis were larger than in

O. julia.

It is interesting to note that some libellulidshave been reported as feeding in

their territories. Amongst these species are included Libellula quadri-

maculata (CORBET, LONGF1ELD & MOORE, 1960), Leucorrhinia

caudalis (PAJUNEN. 1964). L. rubicunda(PAJUNEN, 1966), Nesciothemis

nigeriensis (PARR & PARR, 1974), Orthetrum coerulescens and O.

brunneum(HEYMER, 1969), O. cancellatum (KRONER, 1977)and O. julia

(present study). However, in contrast, other studies have provided evidence

that feeding occurs only before and after territorial periods and never or

rarely at the waterside reproductive site. Species falling into the latter

include Plathemis CAMPANELLA & category lydia (JACOBS, 1955;

WOLF, 1974), Perithemis tenera (JACOBS, 1955), Crocothemis servilia

(HIGASHI, 1969), Orthetrum microstigma, Palpopleura l. lucia, P. l. portia,

Crocothemis Trithemis arteriosa and sp. (GREEN, 1974).

HIGASHI (1969) makes the specific comment that males of Crocothemis

in fields from CORBET servilia tend to feed in the evening away water, and

feed (1962) is of the opinion that male dragonflies do not, as a rule, at the

females. It that there is water, but come to water to meet seems highly likely a

relationship between the length of a particular period of time a libellulid

remains at a waterside territory and whether or not it feeds in that area.

Certainly, of those species regularly feeding in their territories, Libellula

quadrimaculata, Leucorrhinia rubicunda, Nesciothemis nigeriensis, Orthe-

trum cancellatum. O. coerulescens. O. brunneum and O. julia do tend to

remain there for long periods of time. Conversely, non-feeders such as

Plathemis lydia, Perithemis tenera and Palpopleura lucia seem to be species

individual is of duration. where each stay at water relatively short

On the whole, the few detailedstudies of Orthetrumand the closely related

Nesciothemis demonstrate fundamentalbehaviour patterns common to most

of the species. However, each species also shows characteristic and specific

features. information is behavioural When more available, the statement of

HEYMER (1969) ’The two species [ O. coerulescens and O. brunneum] have

developed behaviour patterns fromwhich they can betterand more clearly be

identifiedthan from their morphological systematic characters” will surely be

seen to have general applicability.

ACKNOWLEDGEMENTS

wish with financial I to acknowledge gratefulthanks support made available by the Research

and Publications Committee of the University of Malawi and statistical advice by Prof. J.B.

COPAS, of Salford. wife, made for University My MARION, many helpful suggestions and her

the work 1 assistance in am most grateful. 98 M.J. Parr

REFERENCES

ARAI, Y„ 1972. behaviour Reproductive of Orthetrum albistylum speciosum and O.

trianguläre melania (Libellulidae). Tomho 15 (1-4): 13-17 (In Japanese, with

Engl. s.).

CAMPANELLA, P.J. & L.L. WOLF, 1974. leks Temporal as a mating system in a

temperate zone dragonfly (Odonata: Anisoptera): I: Plathemis lydia (Drury). Behaviour 51 (1-2): 49-87.

CORBET, P S., 1962. A biology of dragonflies. Witherby, London.

CORBET, P S., C. LONGFIELD & N.W. MOORE. I960. Dragonflies. Collins, London.

E. & C. UTZERI, 1974. observations FALCHETT1, Preliminary on the territorial

behaviour of Crocothemis erythraea (Brullé) (Odonata, Libellulidae). Fragm.

em. 10 (3): 295-300.

J., 1974. Territorial behaviour GREEN, in some Nigerian dragonflies. Zoo!. J. Linn. Soc.

55: 225-233.

HASSAN, A.T.. 1978. Reproductive behaviour of Acisoma panorpoides inflatum Selys

(Anisoptera: Libellulidae). Odonaiologica 7 (3); 237-245.

HEYMER, A., 1969. Fortpflanzungsverhalten und Territorialität bei Orthetrum coerulescens

(Fahr.. 1798) und O. brunneum (Fonsc., 1837) (Odonata: Anisoplera). Rev. Comport, animal 3 (2): 1-24.

HIGASHI, K., 1969. Territoriality and dispersal in the population of dragonfly, Croco-

themis servilia Drury (Odonata: Anisoptera). Mem. Fac. Sei. Kyushu Vniv.. (E) 5 (2): 95-113.

ITO, Y., I960. Terrilorialism and residentiality in a dragonfly, Orthetrum albistylum

Uhler. speciosum Ann. em. Soc. Am. 53: 851-853.

JACOBS, M.E.. 1955. Studies on territorialism and sexual selection in dragonflies.

Ecology 36 (4); 566-586.

JOHNSON, C., 1962. A study of territoriality and breeding behaviour in Pachydiplax

longipennis Burmeister (Odonata: Libellulidae). SWesl Nat. 7 (3-4): 191-197.

JOHNSON. C., 1964. The evolution of territoriality in the Odonata. Evolution 18 (I):

89-92.

Notes K.1AUTA, B„ 1964. on some field observations on the behaviour of Leucorrhinia

pectoralis Charp. (Odonata; Libellulidae). Em. Her.. Amst. 24: 82-86.

KORMONDY. E.J., 1961. Territoriality and dispersal in dragonflies (Odonata). Jl. N.Y.

em. Soc. 69: 42-52.

KRONER, U., 1977. Revier- und Fortpflanzungsverhalten von Orthetrum cancellatum

(Linnaeus) (Anisoptera: Libellulidae). Odonaiologica6 (4): 263-270.

LONGFIELD, C., 1955. The Odonata of N. , part I. Publyoes cull. Co. Diam.

Angola 27: 11-64.

MOORE, N.W., 1957. Territory in dragonflies and birds. Bird Study 4 (3): 125-130.

1962a. the of PAJUNEN. V.I., Studies on population ecology Leucorrhinia dubia V.d.Lind.

(Odon., Libellulidae). Ann. zool. Soc. 'Vanamo' 24 (4): 1-79.

PAJUNEN, V.I., 1962b. A description of aggressive behaviour between males of

Leucorrhinia dubia V.d. Lind. (Odon., Libellulidae). Ann. em. fenn. 28: 108-118.

PAJUNEN, V.I., 1963. Reproductive behaviour in Leucorrhinia dubia V.d. Lind, and L.

rubicunda Ann. L. (Odonata, Libellulidae). em. fenn. 29: 106-118.

PAJUNEN, V.I., 1964. Aggressive behaviour in Leucorrhinia caudalis Charp. (Odon.,

Libellulidae). Ann. zool. fenn. I: 357-369.

1966. The PAJUNEN, V.I., influence of population density on the territorial behaviour of

Leucorrhinia rubicunda L. (Odon., Libcllulidae). Ann. zool. fenn. 3: 40-52. Territorial behaviour of Orlhelrum Julia 99

PARR, M.J., 1973. Ecological studies of Ischnura elegans (Odonata: Coenagrionidae).

II. Survivorship, local movements and dispersal. Odonalologica 2 (3): 159-174.

PARR. M.J. & M. PARR, 1974. Studies on the behaviour and ecology of Nesciothemis

nigeriensis Gambles (Anisoptera: Libellulidae). Odonalologica 3 (I): 21-47.

1970. A Orthetrum Occ. PINHEY, E.C.G., new approach to African (Odonata). Pap. natn.

Mus. Rhodesia 4 (30B); 261-321.

M. SAKAGAMI, S.F., H. UBUKATA, IGA & M.J. TODA, 1974. Observations on the

behaviour of some Odonata in the Bonin Islands, with considerations on the

evolution of reproductive behavior in Libellulidae. J. Fac. Sci. Hokkaido Univ.,

(VI), 19 (3): 722-757.

UfiDA, T., 1979. Plasticity of the reproductive behaviour in a dragonfly, Sympetrum parvulum

the males and the Barteneff, with reference to social relationship of density of territories.

Res. Popul. Ecol. 21: 135-152.

YAMAMOTO, Y., 1968. Life-history of the adult dragonfly of Nannophya pygmaea.

Tombo II (3-4): 18-25. (In Japanese, with Engl. s.).